Worktrace

Extract today's Claude Code work history from ~/.claude/history.jsonl and generate a structured markdown summary grouped by project and ticket.

CRITICAL: Execution Rules

MUST DO:

• Run python scripts/worktrace.py --config config.json
• Wait for complete script output before summarizing
• Use the Python script for ALL data extraction

NEVER DO:

• Parse ~/.claude/history.jsonl directly with Bash/jq
• Use find, grep, cat to read session files manually
• Proceed without running the Python script first

WHY: The script handles complex logic (date filtering, ticket extraction, session lookup) that is error-prone and inefficient when done manually with shell commands. The script is optimized for this task.

First-time Setup